The goal of the proposed research is to study functional similarities and differences between proteins with broadly similar structures, but no recognizable sequence similarity. Automated structure comparison programs, together with a new generation of fold prediction algorithms, provide us with many examples of proteins that do, or are likely to, belong to the same structural class. At the same time, such broad structural similarity may or may not be connected to any functional similarity. A database of protein structure and function will be built, allowing a search for groups of proteins with certain structural and functional features. Such groups will then be analyzed to identify structurally and functionally important regions in the sequence and to match them with levels of structural and functional similarity/ divergence. A combination of comparative modeling and structure analysis of protein structures will be used and forged into a standardized set of theoretical tools and procedures that could be predicted using threading and sequence analysis, but whose sequence divergence is large enough to expect significant functional differences. The overall goal of the research described in this project is to bring an understanding of the sequence/structure/function relationship or selected proteins to a level where the detailed predictions of function and mode of action for newly sequenced proteins could be accomplished. With the flood of new sequences coming from large scale genome sequencing projects, it is identifying the possible function of these proteins that would make these data so important.